An Active Appearance Model (AAM) is a computer vision algorithm for matching a statistical model of object shape and appearance with a new image. They are built during a training phase. A set of images, along with the coordinates of the landmarks that appear in all images, is provided to the supervisor.
The model was first presented by Edwards, Cootes and Taylor in a facial analysis at the third International Conference on Face and Gesture Recognition in 1998. Cootes, Edwards and Taylor further described the approach as a general method of computer vision at the European Conference on Computer Vision in the same year. The approach is widely used for face matching and tracking and for the interpretation of medical images.
The algorithm uses the difference between the current estimate of the appearance and the target image to trigger an optimization process. By using the least squares techniques, it can adapt very quickly to new images.
It refers to the active color model (ASM). A disadvantage of ASM is that it only uses shape constraints (analog with some information about the image structure near the landmarks) and does not use all available information – the texture over the target object. This can be modelled with an AAM.
Thank you for visiting.